• UmaMaheswari P
• A.Kumar Kombaiya

Influence maximization has emerged as a crucial problem in the age of ubiquitous digital connection, with social media platforms acting as the main conduits for audience interaction and information dissemination. Predictive accuracy is sometimes hindered by existing models' inability to handle the complexity of heterogeneous networks and successfully integrate multimodal information.In order to increase impact prediction, this study presents a Multi Deep Neural Network (MDNN) architecture that combines DenseNet, Bi-directional CNN (BiCNN), and Enhanced Deep Convolutional Neural Network (Enhanced DCNN). A Biased Bat Algorithm with an Improved Extra Tree Classifier (IETC) optimizes feature selection to preserve just the most discriminative properties, while the preprocessing step uses Biased Renovate K-Means clustering for noise reduction and user segmentation. In contrast to traditional methods, the framework improves generalization and resilience by integrating ensemble learning through the combination of deep neural networks with machine learning classifiers.The suggested model outperforms conventional benchmarks with a maximum prediction accuracy of 98.9%, according to evaluation using datasets from Facebook, YouTube, and Instagram. Based on their different engagement patterns, the results show that Facebook is the most influential platform, followed by YouTube and Instagram. The methodology provides practical insights on the best content tactics, publishing schedules, and engagement techniques in addition to making predictions.The theoretical underpinnings and real-world applications of social media impact maximization are both advanced by this study's production-ready hybrid framework.

Key words: Influence Maximization, Multi Deep Neural Networks, Enhanced DCNN, DenseNet, BiCNN, Biased Renovate K-Means clustering, Biased Bat Algorithm, Ensemble Learning

"Personalized User Recommendations for Social Media Influence Maximization using Enhanced DCNN, Densenet and BiCNN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 10, page no.e666-e674, October-2025, Available :http://www.jetir.org/papers/JETIR2510479.pdf

"Personalized User Recommendations for Social Media Influence Maximization using Enhanced DCNN, Densenet and BiCNN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 10, page no. ppe666-e674, October-2025, Available at : http://www.jetir.org/papers/JETIR2510479.pdf